Identifying Vulnerability Regions of Dust Outbreaks in East Asian Desert Areas: using SMOS, MODIS, and GLDAS

Abstract:

It is now well understood that water, carbon, and energy fluxes at the surface/atmosphere interface are highly dependent on soil moisture (SM). In addition, SM is required to be used as realistic initial states for the SM variables, for climate predictions and weather forecasting. As satellite remote-sensing have developed greatly, global surface SM datasets have been produced based on several satellites. Three satellites-based SM datasets were inter-compared under different land-cover over East Asia to select most reliable satellite for retrieving SM datasets in dust source regions. We estimated satellite sensors with 1) Soil Moisture and Ocean Salinity (SMOS), 2) Advanced Scatterometer (ASCAT), and 3) Advanced Microwave Scanning Radiometer 2 (AMSR2) and Global Land Data Assimilation System (GLDAS) was used as reference datasets. In case of arid areas (desert and semi-desert), SMOS-retrieved SM products showed best accuracy (radiometers have generally exhibited a better performance than scatterometers in dry areas). For this reason, SMOS SM products were utilized to retrieve SM over desert areas. The regions that are susceptible to dust outbreaks were investigated using the dust outbreak probability functions (DOPF). Based on DOPF, about 58% of the total number of dust events occurred in regions with a high level of vulnerability where dust outbreaks were predicted with a probability higher than 60%. The SMOS-based DOPF was calculated to be about 62.4% of the dust outbreak vulnerability (DOV) level of the desert areas. Interestingly, East Asian deserts showed an increasing tendency for a high level of DOV during the study period. Those areas were judged to be sources from which dust could be transported to neighboring countries (e.g., Korea and Japan) which can lead to Asian dust storms. These results may allow us to predict trends of dust outbreaks in order to prepare the corresponding disaster response systems.